{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/causal-unsupervised-semantic-segmentation","title":"Causal Unsupervised Semantic Segmentation","arxiv_id":"2310.07379","date":"2023-10-11","proceeding":null,"authors":["Junho Kim","Byung-Kwan Lee","Yong Man Ro"],"abstract":"Unsupervised semantic segmentation aims to achieve high-quality semantic grouping without human-labeled annotations. With the advent of self-supervised pre-training, various frameworks utilize the pre-trained features to train prediction heads for unsupervised dense prediction. However, a significant challenge in this unsupervised setup is determining the appropriate level of clustering required for segmenting concepts. To address it, we propose a novel framework, CAusal Unsupervised Semantic sEgmentation (CAUSE), which leverages insights from causal inference. Specifically, we bridge intervention-oriented approach (i.e., frontdoor adjustment) to define suitable two-step tasks for unsupervised prediction. The first step involves constructing a concept clusterbook as a mediator, which represents possible concept prototypes at different levels of granularity in a discretized form. Then, the mediator establishes an explicit link to the subsequent concept-wise self-supervised learning for pixel-level grouping. Through extensive experiments and analyses on various datasets, we corroborate the effectiveness of CAUSE and achieve state-of-the-art performance in unsupervised semantic segmentation.","url_abs":"https://arxiv.org/abs/2310.07379v1","url_pdf":"https://arxiv.org/pdf/2310.07379v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"causal-unsupervised-semantic-segmentation","repo_url":"https://github.com/ByungKwanLee/Causal-Unsupervised-Segmentation","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"causal-inference","task_name":"Causal Inference"},{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"self-supervised-learning","task_name":"Self-Supervised Learning"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"unsupervised-semantic-segmentation","task_name":"Unsupervised Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-coco-6","task":"Unsupervised Semantic Segmentation","dataset":"COCO-Stuff-171","model":"CAUSE-TR (ViT-S/8)","rank_in_archive_order":1,"of":4,"metrics":{"Pixel Accuracy":"46.6","mIoU":"15.2"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-coco-7","task":"Unsupervised Semantic Segmentation","dataset":"COCO-Stuff-27","model":"CAUSE (DINOv2, ViT-B/14)","rank_in_archive_order":3,"of":29,"metrics":{"Clustering [Accuracy]":"78.0","Clustering [mIoU]":"45.3"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-coco-7","task":"Unsupervised Semantic Segmentation","dataset":"COCO-Stuff-27","model":"CAUSE (ViT-B/8)","rank_in_archive_order":4,"of":29,"metrics":{"Clustering [Accuracy]":"74.9","Clustering [mIoU]":"41.9"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-coco-8","task":"Unsupervised Semantic Segmentation","dataset":"COCO-Stuff-81","model":"CAUSE-TR (ViT-S/8)","rank_in_archive_order":1,"of":4,"metrics":{"Pixel Accuracy":"75.2","mIoU":"21.2"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-coco-8","task":"Unsupervised Semantic Segmentation","dataset":"COCO-Stuff-81","model":"CAUSE-MLP (ViT-S/8)","rank_in_archive_order":2,"of":4,"metrics":{"Pixel Accuracy":"78.8","mIoU":"19.1"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-pascal-1","task":"Unsupervised Semantic Segmentation","dataset":"PASCAL VOC 2012 val","model":"CAUSE (iBOT, ViT-B/16)","rank_in_archive_order":1,"of":12,"metrics":{"Clustering [mIoU]":"53.4"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-pascal-1","task":"Unsupervised Semantic Segmentation","dataset":"PASCAL VOC 2012 val","model":"CAUSE (ViT-B/8)","rank_in_archive_order":2,"of":12,"metrics":{"Clustering [mIoU]":"53.3"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-semantic-segmentation-on-pascal-1","task":"Unsupervised Semantic Segmentation","dataset":"PASCAL VOC 2012 val","model":"CAUSE (DINOv2, ViT-B/14)","rank_in_archive_order":3,"of":12,"metrics":{"Clustering [mIoU]":"53.2"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2310.07379","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}